Computation of VaR and VaR Contribution in the Vasicek portfolio credit loss model: A comparative study

Abstract

We compare various numerical methods for the estimation of the VaR and the marginal VaR contribution (VaRC) in the Vasicek one-factor portfolio credit loss model. The methods we investigate are the normal approximation, the saddlepoint approximation, a simplified saddlepoint approximation and importance sampling. We investigate each method in terms of speed, accuracy and robustness and in particular explore their abilities of dealing with exposure concentration.